How does spectral morphing work

If im wrong correct me…

It seems like a frequency band splitter combined with a multi fx tool…that allow complete refuckery of a sounds various frequency bands by modulating those various frequency bands to produce new sounds.

Yeah, you pretty much have the concept right.

Spectral morphing just means blending the frequency band characteristics of one sound into another. The changes to the sounds are then reassembled during playback (or whatever kind of execution) into the new sound.

Pretty cool stuff. I love spectral manipulation tools, you can get the weirdest shit out of em.

3 Likes

I’ve always imagined it as multiband splitting on a more quantum scale but I’m not sure if that really matches the reality or not. Even if it does, I’m missing a lot of details in my knowledge regardless. Great explanation, @Guy_Wachtel! I’d really love to get good enough at DSP to make something like this one day, but I’ve got a very long way to go!

2 Likes

This thread reminded me that I haven’t busted out good ol’ MMorph in a few months. Did a little sound design session with it last night. Used Breaktweaker and a Alchemy (made a kind of granular pad thing), then used MMorph and some additional processing. Wild results.

Spectral Morphing is tight

Here’s the chain in you’re interested:

T1: Alchemy Pad
T2: Breaktwerker (lol) with the following plugins in order:

  • MMorph (T1 as sidechain input)
  • Airwindows Isolator2 filter
  • Portal
  • Compressor
  • MAutoDynamicEQ
  • Another compressor
  • MAutopan
  • Logic’s autofilter

Lots and lots of S&H modulation on many of these plugin params.

2 Likes

Wtf is fast fourier transformation in dsp?

I’m gonna try and answer, but my knowledge of this is pretty limited.

I know that the Fourier Transform algorithm essentially takes a sound and translates it from the “time domain” into the “frequency domain” (spectral).

FFT (Fast Fourier Transform) i think is basically another version of the algorithm that does the same thing but more efficiently, as it can do the job in fewer computations.

1 Like

I use Fourier Transform at work on a daily basis.

It is probably the most important computational process in audio.

Guy has the gist, but ill add.

You can use the Fourier Transform to look at an impulse response in the Time Domain or the Frequency Domain. This is an “offline” method and can tell you a great deal of information about what is going on with an audio system. Including room acoustics, time delay, the tone or color of a speaker, how the room is effecting that tone, early reflections, Reverberation Time (aka RT60), Clarity index, the list goes on.

FFT can be used for Real Time Analysis of the audio signal. This is an “online” method, hence the “Fast” part. There are several important and useful applications for this. For instance, you can use it to tune and audio system. Or, you can program a lighting rig to respond to an audio signal in realtime based on spectral cues.

How I use it: I take an impulse response of a speaker, analyse its spectrum, place it against a target curve, compute necessary equalisation parameters, and use the output to equalise said speaker. You can also use another speaker as your target curve. For instance if you have a line array, Ideally you want each box to have an identical frequency response.

Or if you simply just want you speaker to be completely “flat”, you use the inverse of its frequency response and convolve that with the response of the speaker.

Convolution is a whole other can of worms, but it uses the Fourier Transform as well.

3 Likes